Artificial Intelligence Assignment Introduction
AI Course Article
At ZHAW I’m taking a class on artificial intelligence. This blog post is part of the course and mainly discusses the waitbutwhy article which constitutes a mandatory reading for the course.
Types of A.I. Thinking
There are many ways of how a superintelligence can spring into existence as shown with the test-taking-kid analogy in the blog article1, in which the author explains that we might try to write an A.I. fit for taking such a test, or we could make the test take itself. The difference is that some ways of thinking revolve around making A.I.s that help and guide us through our day, whereas another way of thinking is about creating A.I.s that would do everything by themselves and render us obsolete.
Simulate the Brain?
The human brain project showed how difficult it would be to create an A.I. simulating the whole brain, especially considering that right now we barely have the computing power to simulate 1 second of human brain cognition, but once this is possible, an A.I. could operate much more efficiently since our synapsis pathways are limited to 120m/s whereas simulated neurons in an A.I. could virtually travel at the speed of light, rendering thought processes a lot faster, if not better.
A.I. Levels
Artificial Narrow Intelligence
A.k.a. ANI is a limited intelligence able to complete specific tasks efficiently such as playing chess. This has been around for a while now and can be found in a lot of different computer games.
Artificial General Intelligence
A.k.a. AGI is a general intelligence able to operate at a level of a typical human. Experts suggest that it wouldn’t take long going from an AGI to an ASI using recursive improvement methods, e.g. the AGI is tasked with creating a smarter intelligence which in turn is tasked with the same ad infinitum until the result will be an ASI.
Artificial Superintelligence
A.k.a. ASI is a superintelligence able to operate at order of magnitude higher intelligence than humans. There are many possile scenarios and theories how such an intelligence would operate and what would happen to us humans. What people seem to agree on is that the event of such an ASI arising will either be extremely beneficial to us humans or possibly render us extinct. As to the advantages, imagine ASIs bebing able to perform tasks on nanotechnological level such as an ‘eating condom’ whereby smart nanobots in your intestines filter bad food and stop the accumulation of excess fat or nanobots curing damaged cells. On the flip side an ASI might use nanobots to kill us humans in order to free more resources to reach its own goal(s).
My View on A.I.:
Since Elon Musk is an idol for me I’ve always wondered why he is so cautios about A.I., in fact he once called it ‘summoning the demon’. After reading the ‘Wait but Why?’ article on A.I.1 I am aware of a lot of possible dangers, however with our current technology it’s still not possible to render a person immortal and thus we need to die anyway. Why not take the chance and speed up the development of A.I.? We’ll all die eventually, won’t we?